That Favorite Halloween Candy By State Graphic is Worthless Marketing Garbage

Last week, you might have seen a map of our great nation’s favorite Halloween candy by state floating around these here internets. It was put out there by a company called Influenster, and purports to be a survey of people around the country.

It’s garbage. And I know you probably said this when you saw specific things like Candy Corn carrying five different states (the most in the survey) and Toblerone as the choice in Arizona. I’m pretty sure, though, that none of it has validity and so don’t go thinking you are cool, Wisconsin, just because it says you like Laffy Taffy.

If you read the fine print on the document, it says there were 42,238 users. Which sounds like a large enough sample size. And it might be to make a sweeping global characterization of something like “what was the most popular candy in our survey?” (Reese’s Peanut Butter Cups, according to Influenster, and according to Yahoo Parenting, it got 4,348 votes nationwide, just over 10%).

But keep in mind while that’s a large sample size for a binary question, here, respondents were being presented with 30 survey options, according to Yahoo Parenting’s interview with the Influenster CEO. Then, there are 50 states. That means, there were on average, 828 total votes per state (plus D.C.), and when you take the sheer number of brands involved, 28 votes per state per candy would be the average. That’s a lot smaller.

And then again, not all states are created equal. Half of the nation’s population lives in the nine most populous states. Wyoming, for example, is the least populous state and makes up 0.18% of the nation’s population. If this poll reflects that same distribution, then there would have been 77 Wyoming voters on 30 candy brands. The winning candy might have had less than 10 votes. And that winning candy was Candy Corn, by the way.

Even more troubling is this–if there were really 30 options, somehow they all carried at least one state and got to be on that nifty map that is being circulated around. It’s like Oprah in here. You are most popular in a state! You are most popular in a state! Everyone wins a state!

Do you know what the chances of that are? It’s pretty unfathomable to me that if you asked, say, people from every state in the country to identify their favorite actor/actress from a list of 30, all 30 would be the most popular in at least one state. If we assume that the chances of any one candy brand carrying a state is 1 in 30, then there’s a 17.7% chance that any brand you randomly select would carry exactly no states. We would expect 9 of them, on average, to not appear on the map.

Finally, looking at the fine print, only 5% of respondents in this survey were under 18. Meaning there were more old people than actual kids who want candy on Halloween. There was no indication that the adults who voted had children or the poll was limited to such. If we are talking most popular Halloween candy, shouldn’t we actually be looking at who the target audience prefers? I mean, that’s how Toblerone gets a state. How many kids have ever gotten a Toblerone in their bags, ever?

But at least all these brands got their logo circulated, which is what this is really about. I refuse to accept the results as anything close to valid, though. None of my kids voted “Hershey’s Kisses” in Missouri, by the way, and there is a chance I just asked more kids than the survey did in this state.